Flappy machine - Advise needed please!

[cncJim]

Ok so my cnc is generally great (for my very very simple needs!) but it does have 1 major failing that i would like to improve. The gantry is made up from 2 20mm round bars spanning approx. 7-800mm and it supporting a 1.5kW chinese spindle and associated mounts. You can imagine what happens - tons of flex/ruined parts - most noticeable when trying to cut sharp 90 deg turns from y to x, ends up with a lovely big curve if I go to fast! :thumbdown:

Here is a rough diagram.
10489

I also had a situation cutting a long thin pocket running down the x axis - all was going fine until suddenly the spindle seemed to flex down as if was trying to dig into the wood! I guess this was the flexing issue allowing this to happen...

So my question is how could I possibly improve my situation with minimum cost/rework?

Two ideas I thought of are:-
* upgrade the round rail to a larger diameter.
* change the round rail to supported round rail - I was thinking I could grid off part of the support at each end of the rail so it would fit in the existing holes?

Any ideas?

[EddyCurrent]

I suppose you have to first establish exactly where the flexing is taking place, it's maybe not where you first think of.
What about attaching a wide piece of aluminium extrusion or box section from one side piece to the other then fasten supported rail to that.

[cncJim]

I suppose you have to first establish exactly where the flexing is taking place, it's maybe not where you first think of.

Valid point - I admit the only reason for picking on the gantry as being the issue is based on my complex scientific testing - I grabbed the spindle, yanked it around until it flexed! I all feels very rigid except for flexing the spindle up and down.



What about attaching a wide piece of aluminium extrusion or box section from one side piece to the other then fasten supported rail to that.
That could work really well! Thanks for the idea :)

[JAZZCNC]

Jim this is not what your going to want to hear but I'm saying it anyway.!! . . . . Don't bother and start again.

I say this with lots of experience helping folks who took the route your about to take only to regret it later. ONLY out come from taking this route is wasted money and time. Outcome is ALWAYS the same lots of work and still ending with under performing machine.

Honestly don't let others talk you into trying to improve it's false economy that never ends well.

[cncJim]

Thanks for the advice Jazz. I generally agree with what you have said and I can see that my post may have sounded like it came from some blinkered fool blind to reality, but I would say (or hope!) that maybe my situation may be slightly different?....

I can't really start again as I didn't start in the first place! :) I acquired my machine as result of getting carried away on ebay. It was mainly in bits and there was no real info about it but I went for it anyway hoping that if I couldn't get it working I could sell it one for more than I paid.

Lucky for me some cleaning and bolting together was all that was needed :)

I don't think I have the skills (and with 3 kids certainly not the time!) required to put together the kind of machine I would love to own from scratch.

What do I use it for? Well, I have made a nice Aztec jobby for my dad, some vcut signs for friends, a drilling template for my fireplace and thin ply puzzle type toys for my children (I would also like to try some PCBs at some point). I am the first to admit that none of my projects have been perfect (far from it!), but all have been good enough for purpose and for the time and effort I put in.

The only issue I encounter is with the flexing which sometimes costs me time/wood/bits to try and workaround. If I could improve this without spending to much time/effort (£100 and a few hours?) then in my eyes, it would be worth it. But if its not possible to improve then I would still be happy with my machine and continue using it with workarounds.

Anyway, thanks again for replying :)

[JAZZCNC]

I can't really start again as I didn't start in the first place! :) I acquired my machine as result of getting carried away on ebay.

Arghh the dreaded Flee bay strikes again . .Lol . . . .Well in that case then.!!!. . . . . Nothing changes still don't bother and stick it back on Ebay. :whistle:

Without good foundations no amount of money will help and with unsupported rails you have very poor foundations so £100 won't go far or help much at all.


By the way don't think you or anyone else new to building CNC and taking this route are fools.!! . . .I class you and others who have done the same has Maze victims.??? . . . . .You where Duped into a Big Maze with NO exit so end up resorting to digging to get out.!! . . . Not until the problems have been highlighted and explained then decided to ignore sound advise do I consider people fools.!! . . . . . Unfortunately There have been many struck off the list and re-classed has fools.! .:stupid: . . . . . . . . Don't start digging Jim only way is down and don't want to see you on my list of fools either. .:joker:

[cncJim]

Haha! I'll do my best to stay off your list of fools Jazz! No guarantees mind!

Would you mind taking a look at the parts of my machine and give me your opinion? I don't anticipate that I can economically improve the machine but I would like to learn more on what to look for in the future.

My machine's frame is made from approx. 100mm x 40mm welded steel box section
X axis is 1200mm of 25mm unsupported rail
Y axis is 2 x 700mm of 20mm unsupported rail
4 x Zapp steppers
Ballscrews on all
1.5kW Chinese Spindle + Matching VFD
2 x Leadshine SPS705 power supplies
4 x Leadshine Microstep Driver M752
Spindle V3 Board
PCPPS bob

From what I have read on this forum, most of the parts are pretty good (I am only cutting thin wood) but the rails are the big area of issue, mainly from an accuracy point of view. Would that be right?

Also, back to my original issue, does it look like it comes from the flexing on the y axis?

Here is a pic...
10494

If I feed too fast the red points end up being curved rather than a sharp right angle?

I don’t think putting it back on ebay is not an option at the moment though - Not sure how much it would fetch but for the price I paid I wouldn't be able to buy a decent complete machine to replace it. I also really enjoy using it at the moment and feel like I am learning lots every time I fire it up.

Cheers,

Jim

[John McNamara]

Hi Jim

I have used this little program for several years.


You may find the following link useful.....
Straight beams with constant cross sections (Straight beam) (http://www.mitcalc.com/en/pr_beam.htm)


MITCalc is a small Easy to use tool for calculating beam strength and more importantly deflection. It should help you to understand how "floppy" a steel bar really is. Don't confuse strength and deflection. For most machine design deflection is the enemy, as you have found out your 20mm bars easily support the spindle they will never fail in bending. They do however sag and deflect under load.


Many years ago time ago I helped my uncle and a mate restore a 6 cylinder car engine with a cast iron block. we re bored and honed the cylinders to size. While he was testing the bore size he placed a micrometre across the bore longitudinally that is along the block and adjusted it until it just held by its own weight in the bore. He then placed his two index fingers on each side of the particular cylinder and pressed hard. The micrometer fell out. That was my first lesson about steel deflection. a heavy casting is not immune to it.


Get a free trial version of the "beam" module and try it out Only a few Euros if you like it. (I have no connection with the company.)


The program will give the deflection at various points along the length of the beam for a point load, as with a router. This is a static load calculation tool, meaning it does not allow for vibrating or moving loads. Software to do that is not for the feint hearted. My solution is to add a reasonable margin of safety.


Please refer to Beam (structure) - Wikipedia, the free encyclopedia (http://en.wikipedia.org/wiki/Beam_(structure))
Note the different kinds of supports..... Use the fixed setting in MITCalc for firmly attached ends. Although steel fixings also flex a bit!


Experiment with different beam depths and widths. MITCalc will allow you enter your own RHS sections You can also enter your 20 and 25mm bar and compare that to other sections.


Regards
John

[njhussey]

How about some pics of the actual machine to see if it is possible to modify it?

[cncJim]

Thank you John - Very interesting, I will check out MITCalc.

I will upload a pic or two tonight!

[Jonathan]

Can you post a picture of a part which has rounded off corners, which you propose are caused by deflection? The reason I ask is if you have the acceleration set too low for the X and/or Y axis, that will consistently cause the machine to cut noticeable curves in corners if it's in constant velocity mode. Search 'constant velocity' or 'G64' on the forum and you'll find plenty about it. If your acceleration for the Y axis is set too low, then the corners you indicated will be rounded, so I'd start by checking that.

I agree that the list of parts you have are not bad except for the rails, so if you replaced all the rails and make a rigid gantry you'd have a good machine. The 2.2kW spindle would have been nice, due to the larger collets, but that's minor compared to the rails.

With £100 you're struggling, as that's not enough to get rails for both the X and Y axes. Without enough funds to replace every 'weak' part, we need to work out which is the least stiff and replace that as it will have the biggest effect overall. The deflection of unsupported rails for a force applied at the center is proportional to the length cubed, and inversely proportional to the diameter raised to the power 4. That means your X rails will deflect (1200/700)^3*(20/25)^4=2.1 times more the Y rails, so I'd be inclined to replace those first. Having said that, the difference is not that great so you would still notice some gain if you replaced either.

Bear in mind that even if you replaced the rails with just 12mm supported rails, this would still be stronger than your current setup as it is the strength of the supporting structure that matters most when using supported rails, not the rail itself. I would advise using at least 20mm though, especially if wanting to cut harder materials like aluminium, since the extra stiffness of the bearings does help.

Another option would be to invest the £100 in getting the open bearings that the supported rails use (SBR25 & SBR20) and adding your own supports to the existing rails by drilling and tapping holes in them. That could be a courageous decision though...

[JAZZCNC]

Would you mind taking a look at the parts of my machine and give me your opinion? I don't anticipate that I can economically improve the machine but I would like to learn more on what to look for in the future.

Agree with Jonathan and Pretty sure this will be tuning and CV problem. Post the feed rate in G-code or better still the G-code file it's self so we can check.

Regards the rest of machine then agree you have a half decent kit of parts so I'd just look to upgrading the Frame. . BUT. . Don't try to do it on the Cheap because there's very little gain for the Money and time invested.!! . . . Best to Save up and buy decent rails for all axis and Beef up the frame where needed.

Post a picture and lets see what's what.!!

[firetrappe]

I had similar problems with corners cutting as curves. I changed from CV mode to 'Exact Stop' mode in Mach3 and it solved the problem. I know there are probably better ways by tuning CV mode etc, but it might be a quick way of proving whether it's mainly a software or hardware issue.

Si.

[routercnc]

Hi Jim,

Glad to hear you are enjoying the machine despite the problems. The rounded corners sound like the software setting already mentioned, but on the subject of unsupported rails - my first CNC machine used 900mm unsupported 20mm rails for the Y axis (it was a 'landscape' machine rather than the more usual 'portrait' machine) and 600mm unsupported 20mm rails on the X axis. The Z axis was around 300mm unsupported round rail.

I paid about £50 for all the rails and bearings (from a colleague) just to have a play with CNC and see if it was just a passing phase. Well the machine did work, but it flexed badly. The cutter was almost dragged through the work (plywood) at a slight angle so feed rates had to be low. However, I was hooked so I sold the mechanical bits on ebay and designed the machine I have now which is much better (although about to upgrade again!).

I remember playing around with temporary supports clamped to the rail at intervals (if the part being cut was small) but this didn't help alot. As Jazz as mentioned I think you might go the same way - trying bits and pieces but not making too much difference. Jonathan's suggestion might work (change bearings for half-bearings and try to support the rear side of the rail), and I had the same idea with my original machine but didn't do it in the end for fear of grinding the toughened surface to be able to drill into the softer core (to tap) and releasing in-built stresses and distorting the rail. I might have been worrying over nothing there - but it would also have made them more difficult to sell on e-bay after use so I didn't pursue.

About 2 years ago I bought a pair of 16mm supported rails around 1300mm long for ~£100 from Chai on LinearMotionBearings2008 (ebay), so this is probably the best sort of upgrade for your £100 and you could use it on your next (!) machine, but the unsupported Y axis will hold you back.

By the way what rails are on your Z axis because this is also part of the chain?

[EddyCurrent]

The term "Mission creep" springs to mind.

Edit: Jim your original 'rough diagram' did not do your machine justice, it's quite a bit better than that.

[cncJim]

Thank you everyone.

Pics
Mach 3 config:-

http://img.tapatalk.com/d/13/10/25/uqasutuh.jpg
http://img.tapatalk.com/d/13/10/25/3urenyze.jpg
http://img.tapatalk.com/d/13/10/25/3uqe5agu.jpg
http://img.tapatalk.com/d/13/10/25/ydedujur.jpg
http://img.tapatalk.com/d/13/10/25/medeqa5a.jpg
http://img.tapatalk.com/d/13/10/25/ryte5y2e.jpg

Machine pics:-

http://img.tapatalk.com/d/13/10/25/ezurytez.jpg
http://img.tapatalk.com/d/13/10/25/ne7e8uqy.jpg

Love the idea to diy support my existing rails! Not sure I have the tools or skills to pull it off though!

Z axis rails are also 20mm.

Thanks!

[cncJim]

Missed a couple of questions...

I cant post a pic of the part with the rounded corners as it is now installed in my fireplace chamber! (observing the correct distances to combustables, of course!). I can post the g-code used to create it later tonight though.

Jim

[Jonathan]

Your acelleration on X and Y is quite low - you should be fine with about 1000mm/s^2. Try increasing them, or at least set them both to the same thing.

[cncJim]

Thanks Jonathan, I wil try changing the values as suggested.

BTW was that just a general comment on my config or do you think it would be part of my problem?

[JAZZCNC]

Ok well Agree most of this is motor tuning and like Jonathan suggest you balance the tuning up.!! . . . But looking at the frame it has some serious issues.?


Love the idea to diy support my existing rails! Not sure I have the tools or skills to pull it off though

Following on from above statement regards serious issues then I wouldn't suggest you do because there are other problems that will make any attempt pointless without addressing these problems also has they are quite big ones.

The Gantry and the way it fastens to the bearings is a major weakness. It's only mounted to the X axis bearings with 2 bolts per bearings plus the mounting plate area is woe fully inadequate. Also the fastening of the Y axis rails into the puny narrow gantry ends is very weak and the whole thing must flex quite badly and could be some of the problem on direction changes with deep Cuts.

Like said before given the parts you have already and spending a few hundred quid building a new frame will be far far better solution that will give a very good machine that will do every thing you want and to good standard for years, and still be worth more money than you invested when time to sell.!!

[cncJim]

The term "Mission creep" springs to mind.

Edit: Jim your original 'rough diagram' did not do your machine justice, it's quite a bit better than that.

There is always mission creep...... :) though in my experience with IT projects the term "mission creep" is usually applied when the requirement gathering/scoping stage has not been fully completed!

Looking back I guess my diagram is a bit lame!

[cncJim]

Ok well Agree most of this is motor tuning and like Jonathan suggest you balance the tuning up.!! . . . But looking at the frame it has some serious issues.?



Following on from above statement regards serious issues then I wouldn't suggest you do because there are other problems that will make any attempt pointless without addressing these problems also has they are quite big ones.

The Gantry and the way it fastens to the bearings is a major weakness. It's only mounted to the X axis bearings with 2 bolts per bearings plus the mounting plate area is woe fully inadequate. Also the fastening of the Y axis rails into the puny narrow gantry ends is very weak and the whole thing must flex quite badly and could be some of the problem on direction changes with deep Cuts.

Like said before given the parts you have already and spending a few hundred quid building a new frame will be far far better solution that will give a very good machine that will do every thing you want and to good standard for years, and still be worth more money than you invested when time to sell.!!

Thanks for the feedback. I haven't done any kind of proper measuring but if I try and flex the machine at the gantry ends I can see no obvious movement. If I do the same with the spindle, it doesn't take much force before noticable flexing occurs from the 2 unsupported rails.

Must admit I have been browsing the build logs and I can see that I would have a far better machine with a stronger frame.

So I think I have a plan....

I will follow Jazz's advice of starting from scratch - The build logs really are inspirational and I would love solid machine that I can handle all that I throw at it.

BUT

I still investigate a method to stiffen my existing machine as best I can with as little time/effort as possible!

My above plan may seem at odds but I know that starting from scratch will take me alot of time and effort to do right. I think I will not reuse any of my existing machine so I can sell it later on. I will continue to use my machine (it really does work well for the small stuff I am doing) while designing the new one and hopefully some time next year I can start building it!

Off to the build logs to learn....

[JAZZCNC]

I think I will not reuse any of my existing machine so I can sell it later on. I will continue to use my machine (it really does work well for the small stuff I am doing) while designing the new one and hopefully some time next year I can start building it!

Wise choice and one you won't regret.!